You Can’t Speedrun Yourself

Author: Alex Chalmers, for the Cosmos Institute blog URL: https://blog.cosmos-institute.org/p/you-cant-speedrun-yourself Subtitle: The myth of high-level work Published: October 9, 2026 One-line: The optimist’s split between “drudgery” for AI and “good work” for humans doesn’t hold, because the drudgery is often how skill, judgment and attachment get formed. Cheap AI output tempts both people and institutions to skip it, which breeds a modern acedia and hollows out the institutions that train successors.

Key claims

• Acedia is the right frame. The essay opens with Evagrius Ponticus’s “noonday demon,” which made the monk hate “his way of life and his manual work” and long for “an easier, more convenient craft.” Aquinas treated acedia as a sadness that makes spiritual goods feel like burdens. Chalmers argues a modern acedia can attach to any good reachable only through slow, determined work, and that cheap intelligence offers “an approximation of the fruits of determined work” without the work.
• The drudgery/good-work split is harder to draw than optimists think. He cites Satya Nadella’s line that AI will “remove the drudgery of work and unleash creativity,” and the standard routine-vs-judgment guidance from governments and firms. Historically, the burdens technology removed “rarely consisted of pure, unalloyed drudgery”; they were how people acquired skills, habits and attachments.
• Borgmann’s device paradigm. Albert Borgmann’s family stove gave way to central heating, which “procures mere warmth” while removing the attention, skill and seasonal engagement the stove demanded. The stereo did the same to playing music with friends. Borgmann isn’t against boilers or stereos; his point is that “technological disburdenment can remove forms of engagement as well as drudgery.”
• You can only see a practice’s goods from inside it. From MacIntyre’s After Virtue: a child paid in candy to play chess will rationally cheat until they discover goods internal to the game. So outsiders can’t confidently decide which burdens of a practice are dispensable. Entering a practice means accepting “the inadequacy of my own performance” against its standards.
• AI is more radical than Borgmann’s devices. Nobody mistakes a stereo owner for a musician, but a good AI-assisted essay breaks the inference from output to ability. That matters wherever institutions hand out grades, credentials, money and status on the basis of outputs.
• Potemkin institutions. The danger is less cheating than institutions embracing the shortcut, because time saved is legible and lost judgment is not. The example is a junior lawyer’s case research: automating it may keep the output while eroding their grasp of how doctrines fit together. Terence Tao compares AI tools to “taking a helicopter to drop you off at the site.” Once a machine can meet the standard, effortful standards start to feel like arbitrary elite preferences. That is acedia’s rationalization, which John Cassian also described.
• Agency versus understanding. The helicopter does let you visit ten places instead of one. But understanding is what lets practitioners judge plausible answers, and one AI critiquing another draws on similar inherited patterns. Institutions can coast on “inherited standards and machine-assisted outputs for a long time, while spending down the human capital required to preserve and renew standards.”
• The prescription: humility and steadfastness. Few of us have Abbot Paul’s discipline (he wove palm baskets all year and burned them unsold). Like short video against books, AI wins by being easier “in the moments our concentration falters.” Since “there will never be an indisputable taxonomy of formative and nonformative work,” Chalmers says to keep any task that historically formed competent practitioners until there is good reason to think the capability can be built another way. He prefers Benedictine moderation and Evagrius’s hypomonē (“remaining under”) to a “textureless life of want-satisfaction.”

Notable quotes

Cheap intelligence poses a challenge to our longstanding model of formation, in which judgment is acquired through perseverance and apprenticeship.

Instead, he is arguing that technological disburdenment can remove forms of engagement as well as drudgery.

The real issue here is not that it will become easy to cheat. It’s that many institutions will actively embrace the shortcut.

The more completely a machine can produce the visible outputs of a practice, the more irritating and hateful the standards that require real effort on the part of the practitioner.

Institutions can live on inherited standards and machine-assisted outputs for a long time, while spending down the human capital required to preserve and renew standards.

We don’t have to be intellectually persuaded that TikTok is better for us than sustained reading; it need only be easier in the moments our concentration falters.

But if a task was historically part of a route that formed competent practitioners, we should be cautious about removing it until we have a good reason to believe that the same capabilities can be acquired some other way.

How it connects

• Cosmos Institute: publisher, which describes itself as “the Academy for Philosopher-Builders.”
• After Virtue: MacIntyre’s account of practices and internal goods carries the essay’s middle section.
• Apprenticeship: the formation model the essay says cheap intelligence threatens.
• Be Unreasonable: the counterpoint from a practitioner. Thorsten Ball’s advice is to hand agents every task and walk away, which is the “helicopter” taken to its limit.